Prompt Engineering for Writing Revision: EFL Students’ Interactions with ChatGPT-3.5 for Effective Grammar and Reformulation Feedback
JING-XIU SHII, Je-Hyun Kim · The Journal of Human Studies · 2025
This study investigates how Chinese EFL students negotiate prompts with ChatGPT-3.5 to obtain effective writing feedback on grammar correction and text reformulation. Based on qualitative case studies of four exchange students, the research highlights how students’ iterative prompt revisions shape the accuracy and usefulness of AI-generated feedback. Students with higher AI literacy demonstrated greater ability to refine prompts, resulting in feedback that better aligned with their language learning goals, while less experienced users required more attempts to achieve satisfactory results. These findings underscore the importance of explicitly integrating AI literacy instruction―especially prompt engineering―into EFL writing curricula. Educators should equip students with foundational prompting strategies to help them engage more effectively with AI tools like ChatGPT and foster greater learner autonomy in writing development.